What an engagement actually looks like

A B2B SaaS company approaching a priced round usually arrives with three systems telling three different revenue stories. Bookings sit in the CRM, invoicing runs through a billing platform, and the ledger records cash as it lands. Annual recurring revenue at this stage is typically between INR 5 and 15 crore, which is large enough for the gaps between those three sources to matter and small enough that nobody has yet been given the job of closing them.

The common position is a board pack assembled by hand two or three weeks after month end, a revenue figure that cannot be traced to contracts, and an investor request for cohort retention data that has never been produced. The founder is not short of information. What is missing is a single version of it that survives a diligence question.

We begin by reconciling the billing platform to the ledger at contract level, because every later deliverable depends on that tie holding. From there we build a revenue recognition schedule under Ind AS 115 separating recognised revenue, deferred revenue and unbilled receivables, with the treatment stated for each contract type, including annual prepayments, mid-term upgrades and usage-based components.

The first deliverable is that schedule and its supporting reconciliation. The second is a monthly close checklist on a fixed timetable, with responsibilities split by name between the client team and ours, so the close does not depend on any one person being available. The third is a board pack standardised around recognised revenue, net revenue retention, gross margin, burn and runway, with the calculation rule for each metric written down so definitions do not drift between quarters.

Cohort retention is rebuilt from contract-level history rather than inferred from aggregate revenue. Where the underlying history is incomplete, we state the period from which the series is reliable instead of extending it with assumptions.

The engagement is complete when the close runs to a fixed date without our intervention, the billing platform and the ledger reconcile every month, and every figure in the board pack can be traced to a contract or a ledger entry in one step.

The test we hold this to is a diligence test. An investor's analyst should be able to select any revenue number in the pack, ask where it comes from, and be shown the source in the same meeting.

An acquirer evaluating a mid-sized components manufacturer is usually working against a fixed date. The seller has presented reported EBITDA, the growth trend looks orderly, and there are six to eight weeks to a binding offer. Targets of this kind commonly carry related-party rent or management charges, inventory that has not been physically verified for a cycle or two, and a growth story resting on one or two customers.

The common position is that the buyer does not doubt the arithmetic. What is unclear is how much of the reported earnings will still be there after completion, and what a defensible price adjustment looks like if some of it will not.

We begin with a quality of earnings analysis. Reported EBITDA is normalised for related-party rent and management charges, one-off gains and losses, expenditure that has been capitalised, and changes in provisioning policy across the review period. Every adjustment is documented with its source and its rationale, so the buyer can defend or concede each one individually in negotiation rather than arguing a single aggregate.

Working capital is assessed on a twelve-month average basis to set a peg, with the seasonality of the business shown explicitly. Inventory valuation and overhead absorption are tested against production records rather than accepted from the trial balance. Customer concentration is modelled for churn sensitivity, so the buyer can see the earnings profile if the largest account leaves within a year of completion.

Delivery is sequenced to the deal timetable rather than to ours. A red-flag report goes out early enough to change the offer strategy, followed by the full report ahead of the binding date, with a schedule of items sorted by whether they belong in the price, in an indemnity or in escrow.

The engagement is designed to put the buyer into negotiation with a normalised earnings figure they can explain line by line, a working capital peg with a stated basis, and a documented list of issues already allocated between price and contract.

The test we hold this to is whether each adjustment survives challenge from the seller's advisers. An adjustment we cannot evidence is one we should not have made.

A direct-to-consumer brand selling across several marketplaces and its own storefront typically reaches us at the point where reported margin has stopped being believable. Each platform issues settlement reports in its own format on its own cycle, and commission, logistics, returns and penalties are deducted before the payout lands. The finance team, reasonably enough, records the net receipt against the gross invoice.

The common position that follows is that revenue, receivables and margin are all approximate at once, and none of them can be corrected without going back to order level. Deductions applied in error are not disputed, because nobody can demonstrate that they were errors.

We begin at order level. Each platform's settlement report is mapped to ledger entries so that gross sales, commission, logistics charges, return credits and penalties sit in separate accounts instead of arriving as one net figure. That mapping is the reconciliation model, and it is built to be operated by the client's own team rather than by us indefinitely.

The first deliverable is the reconciliation itself, with a documented break resolution process: what counts as a break, who investigates it, how long it may stay open, and the point at which it is raised with the platform as a dispute. We then set a weekly cycle, because a monthly one pushes disputes outside the claim windows the platforms enforce.

Contribution margin reporting by SKU and by channel follows only once the underlying data is clean. Margin analysis built on unreconciled settlements produces confident answers that happen to be wrong, so we do not publish it early.

The engagement is complete when settlements reconcile weekly without our involvement, receivables agree to platform statements at month end, and the brand can state contribution by channel with each deduction visible rather than netted away.

The standard the work is held to is that any charge taken by a platform can be identified, priced and challenged inside that platform's dispute window.

By the time a management team calls us about procurement leakage, they usually have a pattern rather than proof. Maintenance and spares costs at one site run consistently above comparable sites, or a whistleblower message alleges that a vendor is invoicing for work not performed. There is often a name attached, and almost always an internal relationship that makes an ordinary audit awkward.

The common position is that management needs an evidence-led review capable of withstanding scrutiny in a disciplinary process or a court, conducted without alerting the individuals concerned, and completed before the informal version of the story spreads.

We ring-fence the engagement before any testing begins. Access is restricted to a named group, correspondence runs through an agreed channel, and where counsel is involved we settle privilege arrangements at the outset rather than retrofitting them once findings exist.

The work itself covers vendor master analysis for duplicate, dormant and related entities; transaction testing of invoices against work orders, delivery evidence and gate records; trend analysis of unit rates against comparable sites and periods; and review of approval trails for authorisation anomalies such as splitting below a threshold or approvals granted outside the delegation of authority.

Findings are documented to an evidentiary standard. Each one cites the source record, states what it shows and what it does not, and separates what has been established from what remains inference. Where a control failure permitted the transaction, that is recorded separately from any question of who benefited.

The engagement is designed to leave management with a findings report that identifies specific transactions and specific control failures, and that can be passed to counsel, an auditor or a disciplinary panel without rework.

The test we hold this to is that a reader who disagrees with our conclusion can still follow the evidence to it. Anything we cannot support with a source record is reported as unproven and framed as such.

A private company valuing its equity for an ESOP grant and a secondary sale in the same quarter is answering to three audiences with different questions. Option holders want to know what a grant is worth. A secondary buyer wants to know what they are paying for. The auditor wants to know how the figure was arrived at.

The common position is a last priced round that is a year or more old, a business that has changed materially since, and a board that needs a basis it can explain to all three audiences without producing three different answers.

We apply income and market approaches together and state why each is weighted as it is. The discounted cash flow is built on the management plan, with the plan's own assumptions tested and sensitivity ranges run on the variables that actually move the answer rather than on all of them. The comparable companies analysis draws on listed multiples and transaction multiples from the relevant sub-sector, with a written basis for every inclusion and every exclusion.

Adjustments for size, marketability and the preference stack are modelled explicitly rather than absorbed into a single discount. Where a company carries a layered preference structure, the distance between enterprise value and the value of an ordinary share is frequently the whole question, and it is set out on its own.

The report states approach selection, assumptions, sources, adjustments and limitations, and identifies where the purpose requires a registered valuer certificate under Indian law rather than an advisory opinion.

The engagement is complete when the board holds a valuation range whose reasoning is visible, and can explain to an option holder, a buyer and an auditor how the figure was reached from the same document in each conversation.

The standard the work is held to is that a reader can change one assumption, follow the model through, and see the effect. A valuation that cannot be interrogated is of little use to a board.

A multi-facility healthcare group usually approaches us once the receivable balance has grown faster than revenue and no one can say how much of it is real. Insurance and scheme balances age past 180 days, denials are recorded differently at each facility, and resubmission happens when someone remembers rather than when the payer's window requires it.

The common position is two problems running at once: cash that should have been collected and is not being pursued, and a reported receivable that overstates what will ever arrive.

We begin with an ageing and denial diagnostic across every facility, classifying balances by payer, by denial reason and by assessed recoverability. This is done deliberately before any collections activity, because pursuing the whole balance with equal effort is how groups in this position stay stuck.

From there we introduce a standard follow-up cadence with defined escalation steps and a single owner for each payer relationship, together with a resubmission workflow keyed to each payer's own timelines rather than to internal convenience. A weekly collections review works the classified list in priority order, and every action is recorded against the account, so the client sees what has been done rather than what was intended.

Provisioning recommendations are prepared for balances assessed as unrecoverable, with the basis stated for each category, so that the reported receivable reflects economic reality and collections effort is directed at what remains.

The engagement is designed to leave the group with a payer-level view of collectability, a follow-up cycle that runs to a cadence rather than to memory, and a documented provisioning basis the auditor can review.

The test we hold this to is whether any balance on the ledger can be placed on request into one of three states: actively pursued, in dispute with a stated next step, or provided against for a stated reason.